# Vibe Gain: How AI Unlocked Hidden Coding Time

*Interactive version: [View on web](/articles/os-vibe-gains.html)*


Analysis of [@restlessronin's GitHub](https://github.com/restlessronin) activity shows 2-2.5x productivity gains from AI collaboration. The data reveals significant shifts in coding patterns—more frequent sessions, tighter iteration cycles, sustained daily output.

@restlessronin's experience: AI eliminated activation energy across all coding activities, transforming fragmented time into productive sessions.

> **Context:** @restlessronin works part-time on pro-bono and open source projects, fitting coding sessions around other commitments.

This analysis examines only commit patterns—no PRs, issues, or code quality metrics—revealing what's possible from timestamp data alone.

## From Sporadic to Sustained

The transformation is instantly visible in these commit timelines. <span class="period-pre-ai">Pre-AI</span> development shows feast-or-famine patterns: intense bursts separated by gaps. With <span class="period-recent-ai">AI collaboration</span>, it's a steady flow of daily contributions. Each mark represents a commit, with colors and shapes indicating repositories and opacity showing lines changed.

*Interactive commit activity visualization - view at: /articles/os-vibe-gains.html*


*Interactive commit activity visualization - view at: /articles/os-vibe-gains.html*


Looking at hourly distribution reveals how work spread across the day. Pre-AI shows irregular peaks and valleys—some hours see 11% of daily commits, others just 2.5%. With AI, most working hours see 4-8% of commits, with notable evening concentration (12% at hour 20).

Day boundaries are centered in the automatically determined low-activity periods shown in this histogram.

## Commits by hour of day (percentage %)

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 0:00-1:00 | 2.00 | 0.00 |
| 1:00-2:00 | 0.00 | 0.00 |
| 2:00-3:00 | 1.00 | 4.00 |
| 3:00-4:00 | 3.00 | 10.00 |
| 4:00-5:00 | 12.00 | 25.00 |
| 5:00-6:00 | 26.00 | 67.00 |
| 6:00-7:00 | 45.00 | 75.00 |
| 7:00-8:00 | 40.00 | 54.00 |
| 8:00-9:00 | 39.00 | 70.00 |
| 9:00-10:00 | 54.00 | 74.00 |
| 10:00-11:00 | 54.00 | 64.00 |
| 11:00-12:00 | 37.00 | 53.00 |
| 12:00-13:00 | 12.00 | 62.00 |
| 13:00-14:00 | 23.00 | 64.00 |
| 14:00-15:00 | 13.00 | 47.00 |
| 15:00-16:00 | 45.00 | 119.00 |
| 16:00-17:00 | 33.00 | 95.00 |
| 17:00-18:00 | 19.00 | 55.00 |
| 18:00-19:00 | 13.00 | 41.00 |
| 19:00-20:00 | 8.00 | 10.00 |
| 20:00-21:00 | 4.00 | 2.00 |
| 21:00-22:00 | 0.00 | 1.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



## The Numbers Tell the Story

These visual patterns translate to concrete metrics. First, the six-month totals:

**Six-Month Comparison**

| Metric        | Pre-AI | Recent-AI | Change |
| ------------- | -----: | --------: | -----: |
| Total days    |    183 |       181 |  -1.1% |
| Active days   |     99 |       145 |   +46% |
| Total commits |    483 |       992 |  +104% |

## Daily Velocity Doubled

The foundation of these gains lies in daily productivity, which is based on the natural day boundaries from the hour-of-day patterns.

The output per active day is markedly higher. Median daily commits doubled from 3 to 6. The band of highly productive days (5-19 commits) increased from 34% to 55% of all active days.

## Code commits per active day

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 3.00 | 6.00 |
| Min | 1.00 | 1.00 |
| Max | 21.00 | 32.00 |
| Q1 | 2.00 | 3.00 |
| Q3 | 7.00 | 10.00 |
| P5 | 1.00 | 1.00 |
| P95 | 13.00 | 17.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 22.00 | 18.00 |
| 2 | 16.00 | 13.00 |
| 3-4 | 25.00 | 30.00 |
| 5-9 | 21.00 | 47.00 |
| 10-19 | 13.00 | 32.00 |
| 20-49 | 2.00 | 5.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



This wasn't just more frequent small commits—the scale of daily work grew. Lines of code show similar gains. Days with 200-999 lines grew from 33% to 46% of active days.

## Lines changed per active day

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 238.00 | 532.00 |
| Min | 1.00 | 1.00 |
| Max | 10010.00 | 9620.00 |
| Q1 | 52.00 | 217.00 |
| Q3 | 871.00 | 1218.00 |
| P5 | 3.00 | 15.00 |
| P95 | 5020.00 | 3893.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 1.00 | 1.00 |
| 2-4 | 6.00 | 1.00 |
| 5-9 | 1.00 | 2.00 |
| 10-24 | 6.00 | 7.00 |
| 25-49 | 10.00 | 7.00 |
| 50-99 | 13.00 | 3.00 |
| 100-199 | 8.00 | 12.00 |
| 200-499 | 19.00 | 39.00 |
| 500-999 | 14.00 | 28.00 |
| 1000-1999 | 9.00 | 26.00 |
| 2000-4999 | 6.00 | 14.00 |
| 5000-9999 | 5.00 | 5.00 |
| 10000+ | 1.00 | - |

*View interactive chart at: /articles/os-vibe-gains.html*

---



## Commits Stayed Meaningful

A natural concern: did doubling commit frequency mean each commit became less substantial? Higher frequency could mean smaller, less meaningful commits. The data shows otherwise.

Lines per commit distributions retained their fundamental shape while shifting modestly upward—median increased from 18 to 22, with fewer very small commits and more mid-sized changes.

## Lines changed per commit

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 20.00 | 22.00 |
| Min | 1.00 | 1.00 |
| Max | 7476.00 | 7494.00 |
| Q1 | 4.00 | 6.00 |
| Q3 | 103.00 | 116.00 |
| P5 | 2.00 | 2.00 |
| P95 | 852.00 | 678.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 14.00 | 28.00 |
| 2-4 | 104.00 | 177.00 |
| 5-9 | 59.00 | 112.00 |
| 10-24 | 74.00 | 191.00 |
| 25-49 | 56.00 | 88.00 |
| 50-99 | 38.00 | 123.00 |
| 100-199 | 53.00 | 89.00 |
| 200-499 | 29.00 | 100.00 |
| 500-999 | 18.00 | 45.00 |
| 1000-1999 | 9.00 | 25.00 |
| 2000-4999 | 9.00 | 8.00 |
| 5000-9999 | 3.00 | 1.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



In fact, commit scope actually broadened. Single-file commits dropped from 64% to 48%, replaced by multi-file changes.

## Files changed per commit

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 1.00 | 2.00 |
| Min | 1.00 | 1.00 |
| Max | 37.00 | 84.00 |
| Q1 | 1.00 | 1.00 |
| Q3 | 2.00 | 3.00 |
| P5 | 1.00 | 1.00 |
| P95 | 11.00 | 10.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 308.00 | 472.00 |
| 2 | 72.00 | 191.00 |
| 3-4 | 42.00 | 163.00 |
| 5-9 | 27.00 | 111.00 |
| 10-19 | 24.00 | 43.00 |
| 20-49 | 6.00 | 10.00 |
| 50+ | - | 2.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



## Anatomy of 2x Productivity

The timing data reveals significant pattern changes: more frequent context switches, tighter iteration cycles, and higher session productivity.

### Higher Hourly Velocity

Starting at the hour level: both time and output increased. Active coding hours per day increased 50% (median: 2 to 3). Days with just 1-2 hours became less common, replaced by 3-5 hour days.

## Active coding hours per day

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 2.00 | 3.00 |
| Min | 1.00 | 1.00 |
| Max | 9.00 | 12.00 |
| Q1 | 1.00 | 2.00 |
| Q3 | 4.00 | 5.00 |
| P5 | 1.00 | 1.00 |
| P95 | 6.00 | 7.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 35.00 | 32.00 |
| 2 | 22.00 | 22.00 |
| 3 | 16.00 | 27.00 |
| 4 | 10.00 | 23.00 |
| 5 | 7.00 | 17.00 |
| 6 | 4.00 | 10.00 |
| 7 | 3.00 | 10.00 |
| 8 | 1.00 | 2.00 |
| 9 | 1.00 | 1.00 |
| 10 | - | - |
| 11 | - | - |
| 12 | - | 1.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



More importantly, each hour became more productive. Within those hours, velocity doubled. Median commits per hour went from 1 to 2.

## Commits per hour

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 1.00 | 2.00 |
| Min | 1.00 | 1.00 |
| Max | 7.00 | 9.00 |
| Q1 | 1.00 | 1.00 |
| Q3 | 2.00 | 3.00 |
| P5 | 1.00 | 1.00 |
| P95 | 4.00 | 4.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 136.00 | 237.00 |
| 2 | 76.00 | 131.00 |
| 3-4 | 45.00 | 112.00 |
| 5-9 | 7.00 | 21.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



Lines per hour followed the same pattern, with median output increasing from 52 to 91 lines—a 75% gain that mirrors the commit frequency improvement.

## Lines per hour

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 59.00 | 86.00 |
| Min | 1.00 | 1.00 |
| Max | 7786.00 | 7494.00 |
| Q1 | 15.00 | 20.00 |
| Q3 | 187.00 | 290.00 |
| P5 | 2.00 | 2.00 |
| P95 | 1548.00 | 1414.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 4.00 | 13.00 |
| 2-4 | 23.00 | 35.00 |
| 5-9 | 20.00 | 34.00 |
| 10-24 | 39.00 | 61.00 |
| 25-49 | 32.00 | 59.00 |
| 50-99 | 37.00 | 59.00 |
| 100-199 | 41.00 | 72.00 |
| 200-499 | 24.00 | 84.00 |
| 500-999 | 19.00 | 40.00 |
| 1000-1999 | 10.00 | 29.00 |
| 2000-4999 | 8.00 | 12.00 |
| 5000-9999 | 3.00 | 1.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



### More Frequent Sessions

Drilling down further reveals how those hours filled with more activity. Threshold analysis of commit intervals identified natural session boundaries—gaps of 96 minutes in the pre-AI period versus 81 minutes with AI collaboration.

## Time between commits

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 15.18 | 10.88 |
| Min | 0.15 | 0.10 |
| Max | 658.95 | 665.92 |
| Q1 | 4.23 | 2.73 |
| Q3 | 78.17 | 54.05 |
| P5 | 0.40 | 0.43 |
| P95 | 348.40 | 245.80 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| < 20 sec | 8.00 | 28.00 |
| 20-60 sec | 38.00 | 107.00 |
| 1-2 min | 22.00 | 69.00 |
| 2-5 min | 45.00 | 136.00 |
| 5-10 min | 43.00 | 124.00 |
| 10-15 min | 34.00 | 80.00 |
| 15-20 min | 23.00 | 52.00 |
| 20-30 min | 25.00 | 52.00 |
| 30-45 min | 24.00 | 55.00 |
| 45-60 min | 15.00 | 28.00 |
| 60-90 min | 17.00 | 52.00 |
| 90-120 min | 11.00 | 28.00 |
| 2-3 hours | 27.00 | 56.00 |
| 3-4 hours | 13.00 | 43.00 |
| 4-6 hours | 20.00 | 32.00 |
| 6-8 hours | 11.00 | 18.00 |
| 8-12 hours | 8.00 | 5.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



Using these thresholds to identify individual sessions allowed us to zoom in on daily work patterns. The distribution shifted from 1-2 sessions to 3-4 sessions per day.

## Coding sessions per active day

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 2.00 | 2.00 |
| Min | 1.00 | 1.00 |
| Max | 4.00 | 6.00 |
| Q1 | 1.00 | 1.00 |
| Q3 | 2.00 | 3.00 |
| P5 | 1.00 | 1.00 |
| P95 | 3.00 | 4.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 42.00 | 48.00 |
| 2 | 33.00 | 35.00 |
| 3 | 19.00 | 36.00 |
| 4 | 5.00 | 20.00 |
| 5 | - | 4.00 |
| 6 | - | 4.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



### Consistent Sessions

Median session duration increased from 25 to 31 minutes, but the overall distribution shape remained consistent.

## Coding session duration

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 25.67 | 27.72 |
| Min | 0.28 | 0.20 |
| Max | 314.13 | 223.97 |
| Q1 | 6.63 | 7.17 |
| Q3 | 61.30 | 64.32 |
| P5 | 0.40 | 0.50 |
| P95 | 184.85 | 148.63 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| < 20 sec | 1.00 | 4.00 |
| 20-60 sec | 10.00 | 14.00 |
| 1-2 min | 1.00 | 13.00 |
| 2-5 min | 7.00 | 18.00 |
| 5-10 min | 15.00 | 16.00 |
| 10-15 min | 7.00 | 16.00 |
| 15-20 min | 7.00 | 19.00 |
| 20-30 min | 7.00 | 16.00 |
| 30-45 min | 16.00 | 29.00 |
| 45-60 min | 9.00 | 23.00 |
| 60-90 min | 7.00 | 22.00 |
| 90-120 min | 11.00 | 13.00 |
| 2-3 hours | 3.00 | 19.00 |
| 3-4 hours | 6.00 | 6.00 |
| 4-6 hours | 1.00 | - |

*View interactive chart at: /articles/os-vibe-gains.html*

---



The time between sessions tells the rest of the story. Breaks between sessions shortened from 227 to 184 minutes median—quicker returns to code.

## Time between coding sessions

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 224.93 | 173.42 |
| Min | 97.13 | 79.18 |
| Max | 658.95 | 665.92 |
| Q1 | 156.50 | 123.72 |
| Q3 | 341.08 | 242.25 |
| P5 | 109.55 | 82.93 |
| P95 | 509.20 | 440.18 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| < 20 sec | - | - |
| 20-60 sec | - | - |
| 1-2 min | - | - |
| 2-5 min | - | - |
| 5-10 min | - | - |
| 10-15 min | - | - |
| 15-20 min | - | - |
| 20-30 min | - | - |
| 30-45 min | - | - |
| 45-60 min | - | - |
| 60-90 min | - | 21.00 |
| 90-120 min | 7.00 | 28.00 |
| 2-3 hours | 27.00 | 56.00 |
| 3-4 hours | 13.00 | 43.00 |
| 4-6 hours | 20.00 | 32.00 |
| 6-8 hours | 11.00 | 18.00 |
| 8-12 hours | 8.00 | 5.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



### Faster Iteration

Within each session, the development cycle itself accelerated. Gaps under five minutes grew from 38% to 45% of all intra-session intervals.

## Time between commits within sessions

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 8.78 | 6.42 |
| Min | 0.15 | 0.10 |
| Max | 96.23 | 78.17 |
| Q1 | 2.52 | 1.78 |
| Q3 | 24.55 | 17.28 |
| P5 | 0.37 | 0.37 |
| P95 | 64.88 | 53.73 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| < 20 sec | 8.00 | 28.00 |
| 20-60 sec | 38.00 | 107.00 |
| 1-2 min | 22.00 | 69.00 |
| 2-5 min | 45.00 | 136.00 |
| 5-10 min | 43.00 | 124.00 |
| 10-15 min | 34.00 | 80.00 |
| 15-20 min | 23.00 | 52.00 |
| 20-30 min | 25.00 | 52.00 |
| 30-45 min | 24.00 | 55.00 |
| 45-60 min | 15.00 | 28.00 |
| 60-90 min | 17.00 | 31.00 |
| 90-120 min | 4.00 | - |

*View interactive chart at: /articles/os-vibe-gains.html*

---



This faster iteration translated directly to higher session output. Single-commit sessions dropped from 42% to 36%, replaced by more multi-commit bursts (3-4 commits rising from 20% to 26% of sessions).

## Commits per session

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 2.00 | 2.00 |
| Min | 1.00 | 1.00 |
| Max | 16.00 | 17.00 |
| Q1 | 1.00 | 1.00 |
| Q3 | 3.00 | 4.00 |
| P5 | 1.00 | 1.00 |
| P95 | 7.00 | 8.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 77.00 | 121.00 |
| 2 | 46.00 | 69.00 |
| 3-4 | 36.00 | 89.00 |
| 5-9 | 22.00 | 51.00 |
| 10-19 | 4.00 | 10.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



Lines per session increased 63% (96 to 156 median), indicating higher productivity within sessions.

## Lines per session

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 96.00 | 155.00 |
| Min | 1.00 | 1.00 |
| Max | 8234.00 | 7494.00 |
| Q1 | 27.00 | 34.00 |
| Q3 | 334.00 | 487.00 |
| P5 | 3.00 | 2.00 |
| P95 | 1995.00 | 1869.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 4.00 | 9.00 |
| 2-4 | 10.00 | 19.00 |
| 5-9 | 10.00 | 12.00 |
| 10-24 | 19.00 | 28.00 |
| 25-49 | 24.00 | 39.00 |
| 50-99 | 26.00 | 34.00 |
| 100-199 | 33.00 | 50.00 |
| 200-499 | 17.00 | 63.00 |
| 500-999 | 18.00 | 37.00 |
| 1000-1999 | 11.00 | 31.00 |
| 2000-4999 | 5.00 | 14.00 |
| 5000-9999 | 5.00 | 2.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



## Deeper Project Involvement

This productivity increase wasn't just busywork—it enabled substantially deeper project involvement, with sustained work on multiple complex projects rather than scattered contributions.

Daily repository engagement increased slightly—median remained at 1 repository per day, but multi-repository days became more common.

## Repositories per active day

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 1.00 | 1.00 |
| Min | 1.00 | 1.00 |
| Max | 4.00 | 4.00 |
| Q1 | 1.00 | 1.00 |
| Q3 | 1.00 | 2.00 |
| P5 | 1.00 | 1.00 |
| P95 | 2.00 | 3.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 77.00 | 93.00 |
| 2 | 17.00 | 39.00 |
| 3 | 4.00 | 14.00 |
| 4 | 1.00 | 1.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



During the AI period, work concentrated on two well-received open-source repos, this website, and several pro bono projects—all initiated after the pre-AI period. @restlessronin attributes this additional capacity to AI enabling complex initiatives that weren't previously practical.

Deep engagement increased. Repositories with over 100 commits grew from 1 to 4 projects (6% to 21% of all repositories).

## Commits per repository

| Metric | Pre-AI | Recent-AI |
|--------|--------|-----------|
| Median | 6.00 | 19.00 |
| Min | 1.00 | 1.00 |
| Max | 146.00 | 281.00 |
| Q1 | 2.00 | 3.00 |
| Q3 | 53.00 | 55.00 |
| P5 | 1.00 | 1.00 |
| P95 | 99.00 | 258.00 |

| Bin | Pre-AI | Recent-AI |
|-----|--------|-----------|
| 1 | 1.00 | 2.00 |
| 2 | 4.00 | 2.00 |
| 3-5 | 1.00 | 5.00 |
| 6-10 | 3.00 | - |
| 11-20 | 1.00 | 2.00 |
| 21-40 | - | - |
| 41-100 | 4.00 | 4.00 |
| 100+ | 1.00 | 4.00 |

*View interactive chart at: /articles/os-vibe-gains.html*

---



## What This Means

**Pattern Changes:** The data shows more sessions per day, shorter gaps between sessions, and higher productivity within sessions. These patterns suggest reduced barriers to starting and resuming coding sessions.

**Developer Experience:** The consistent experience was effortless transitions—starting sessions, switching between issues, getting unstuck, iterating, and moving between repositories felt immediate. Tackling ambitious new projects became as natural as quick fixes. Time slots of any length became immediately productive.

AI collaboration enables instant productivity—whether in brief windows between meetings, extended work sessions, or when switching between different codebases and problems.

## Beyond One Developer

This analysis reveals patterns in a single developer's workflow. Do these gains generalize?

We extended this analysis to 18 prominent open-source developers—the creators of languages, frameworks, and tools used by millions. The [10x in the AI Age](10x-in-the-ai-age.html) article explores their patterns, finding similar transformations: some developers dramatically increased output, others improved efficiency, and many showed structural shifts in how they work.

Want to investigate the data yourself? The [Vibe Gain Explorer](vibe-gain-explorer.html) provides interactive visualizations of all 18 developers' timelines, daily patterns, and productivity metrics—the same analysis framework used here, applied across elite developers working on everything from Linux to Vue.js to RabbitMQ.

## Methodology

**Comparison periods:** June-November 2022 (pre-AI) and November 2024-April 2025 (mature human-in-the-loop AI collaboration, before agentic tools became widespread), chosen for consistent activity levels.

**Data Source:** All accessible GitHub repositories. Documentation-only commits to www.cyberchitta.cc were excluded from productivity metrics but retained for session boundaries.

**AI Collaboration:** All AI interactions via [chat interfaces](llm-ctx-why.html) (no IDE integrations like Cursor, Windsurf, or copilots). Primary collaboration with Claude Sonnet 3.5-3.7, additional work with Grok-3 and Gemini 2.0 Pro. All interactions used [llm-context](https://github.com/cyberchitta/llm-context.py) for [complete project awareness](full-context-magic.html).

**Analysis:** [vibe-gain](https://github.com/cyberchitta/vibe-gain) visualization library.

**Limitations:** Session durations are lower bounds (first to last commit). Pre-/post-commit work not captured.

## Credits

Original concept and project vision by @restlessronin. Analytical framework brainstormed with @grok-3.

Article drafted by @claude-4-opus, extensively copy-edited by @claude-4-sonnet, reviewed, restructured and revised by @grok-4 and @gemini-2.5-pro, finished by @claude-4-opus.

Data visualization and interactive chart code created by @claude-3.7-sonnet and @claude-4-sonnet.

Showrunner: @restlessronin


---

## Document History

**Jan 15, 2026:** Added cross-references to 10x and explorer articles, refined methodology description

**Dec 31, 2025:** Matched text to stats output by latest vibe-gain.

**Aug 4, 2025:** Added day boundary. Improved flow and clarity.

**Aug 3, 2025:** Improved timeline plot visuals.

**Jul 22, 2025:** Fixed data accuracy and improved clarity.

**Jul 21, 2025:** Moved repository section for better narrative flow. Updated metrics.

**Jul 20, 2025:** Eliminated UX stutter during layout/resize/rotation.

**Jul 19, 2025:** Added strip plot sessions.

**Jul 18, 2025:** Added hourly lines analysis.

**Jul 17, 2025:** Added box-plot values legend.

**Jul 16, 2025:** Added box-plot stats legend.

**Jul 15, 2025:** Separated interpretation from observations for clearer analysis

**Jul 14, 2025:** Refined charts and copy-edit. Changed title.

**Jul 13, 2025:** Added hourly distributions and session productivity metrics.